Project Background: We have reported that chronic kidney disease (CKD) occurs in about one third of all veterans with diabetes (DM-CKD). We also published that consistency of nephrologist care over one year was associated with reduced risk of dialysis-free death independent of other ongoing clinical care. Additionally, veterans with DM-CKD are frequently hospitalized for ambulatory care sensitive conditions which could be prevented with timely and effective outpatient care prior to hospitalization. These have been defined by Agency for Health Research and Quality (AHRQ) as Prevention Quality Indicators (PQIs). However, whether consistency of nephrologist care is associated with decreased risks of PQIs is unknown. Understanding of benefits of subspecialty care independent of primary care in reducing risks of PQIs, as well as identification of individual level characteristics that predict risks for PQIs hav implications for improving care for this high risk population and may inform system redesign for subspecialty care within the Patient Aligned Care Team for veteran patients with DM -CKD. Project Objectives: We propose to study rates of PQIs and ambulatory care in Veteran Health Administration (VHA) patients with diabetes and CKD (DM-CKD). Among patients with DM-CKD, we proposed the following specific aims. (1) Study rates and trends of nine currently defined AHRQ PQI hospitalizations and three proposed PQIs; evaluate factors associated with risks of PQIs. (2) Study utilization of subspecialty (nephrology, endocrinology, and cardiology) and primary care; evaluate factors associated with use and consistency of subspecialty care. (3) Evaluate whether use and consistency of subspecialty care is associated with risk of PQIs; and (4) Develop predictive models for PQI hospitalizations as a potential tool for DM-CKD clinical management. Project Methods: We will use an established research database, the Diabetes Epidemiology Cohorts (DEpiC). Among patients with diabetes, we will establish a study cohort of patients in FY 2000-2006 with CKD, stage 2-4 (estimated glomerular filtration rate (eGFR) of 15-90 ml/min/1.73m2) and follow them for PQIs through FY 2009. Primary study outcomes are individual as well as composite PQIs. Composite PQIs include metabolic decompensation, lower extremity amputations/ulcers, cardiovascular events, and infections (pneumonia and urinary tract infection). Key variables are subspecialty care in nephrology, endocrinology, and cardiology. Other covariates include patient-level factors (e.g., demographics, co- morbid conditions, renal status, prior hospitalization, VHA-Medicare dual system utilization, rurality of residence) and selected VHA facility-level factors (availability of subspecialists). For data analysis, crude and age-sex standardized annual rates of PQIs and presence of subspecialty care will be calculated. We will evaluate their annual trends using logistic regressions with Generalized Estimating Equations (GEEs). Cox proportional hazards models will be used to evaluate the associations of consistency of subspecialty care and other independent variables upon risks of PQIs. Predictive models for risks of PQIs will be developed and validated for model predictability.